steve1977

I had 2 ZFS disks in a mirrored pool. I removed the pool and then tried to mount one of the disks in an array. Unfortunately, it gives me an error message that the disk is unmountable. I then tried to mount it again as ZFS pool, but this also gives me the same error. Any idea whether I can get the content back? Attached diagnostic file. Thanks! tower-diagnostics-20260201-1526.zip

How to "break" the mirror. After breaking, can I mount both as UD devices and can access the same data on both UD devices?

Any thoughts?

I have a ZFS pool set up (one disk mirrored to another one). While I am happy with it, I am reconsidering and move (back) to a traditional Unraid Array with XFS disks (1 disk with 1 parity disk). I am wondering what's the fastest & safest way to convert my ZFS mirrored pool to a 1 disk / 1 parity XFS array.

I will now need to decide going with an Unraid XFS Array or with a ZFS Pool. For my 1 disk / 1 mirror setup, is there any reason I should go with Array over Pool? Or a clear advantage of an ZFS Pool?

Does it make a difference where to attach the flash disk to? USB 2.0 USB 3.2 Gen 1 USB 3.2 Gen 2

I just changed the USB port of the enclosure and this led to a the flash disk shows as corrupted. I remember I had this before when I plug in/out USB devices. After restarting Unraid, all is normal again. Also when this happens, the GUI crashes and is no longer accessible. I can still access docker GUI and VMs though. Does this point to any issue? I can change the flash disk if you think that's the reason?

It is a regular Unraid install. Just using the "USB Manager" plugin. I uninstalled the plugin, but no change. Somehow not identifying the enclosure. Quite weird as the enclosure is being identief on my other Unraid server. Any idea what else I can do to trouble-shoot. Let me try a few other USB ports.

Got it. Will keep 580.119.02 for my 1060. For the other server with 2 GPUs, it now worked out. It only shows in the plugin one of my 2 GPUs though. Is it meant to only support one?

Enclosure & disk work with another PC. I just tried a different USB port. Also doesn't seem to work. Sending another diagnostic log. Thanks for your help! Also noticed that I have a tab "USB", which gives me below info. Maybe that can help lead to the solution? tower-diagnostics-20260110-1041.zip

Got it, so this is expected behaviour and still can access the files via /mnt/user/..., but just "ls" doesn't work for exclusive shares?

Here we go: root@Tower:~# ls -l /mnt/user total 8 lrwxrwxrwx 17 nobody users 4096 Jan 10 07:56 appdata -> ../cache/appdata/ drwxrwxrwx 1 myname users 149 Jan 10 03:07 backup/ drwxrwxrwx 1 myname users 35 Jan 10 08:29 backup-music/ drwxrwxrwx 1 nobody users 63 Jan 9 23:55 data/ lrwxrwxrwx 3 nobody users 24 Dec 20 2022 domains -> ../cache/domains/ lrwxrwxrwx 2 nobody users 106 Jul 3 2025 isos -> ../cache/isos/ lrwxrwxrwx 4 nobody users 56 Dec 25 2022 system -> ../cache/system/ root@Tower:~#

Got it. Then this option is out. So, what about XFS Array vs ZFS Pool. Given only one disk with mirror/parity, many of the XFS array advantages do not apply. Anything I am missing why XFS array may still be superior? Or no brainer to go with a ZFS pool?

Now that Unraid can run several pools and can operate without an array, I have a few questions. I have two Unraid servers. One has my major media storage, which runs as XFS Array. I love it. It allows different size and limits disk loss risk to 1 disk evenif parity and 1 disk die at the same time. No plan to change this and it works very stable. I have a second Unraid server. I will only have one disk (in addition to the cache disk) for my music library (for Plex / UPNP) and personal backups. Most important factor will be safety, whiled speed plays some role for the music files. It's many small files as you can imagine. It's anyways a sata spinner, so won't be lightning fast. This second server is currently set up as an array with one XFS disk and a second XFS disk for parity. I am thinking to change this to a ZFS array or a ZFS pool. Would love to hear any view whether there is any downside of doing so? And the advantages of ZFS array vs ZFS pool in this context? Thanks all for your help!

Rebooted and still cannot access appdata as user share. See below diagnostic file. Thanks for looking into this. tower-diagnostics-20260110-0937.zip

Thanks for the development of this plugin. I have two unraid servers, for which I'd like to use the plugin. 1) Unraid Server 1 Plugin installed with Nvidia 1060 with driver 580.119.02 Possible to upgrade driver or that's as good as it gets with this old card? 2) Unraid Server 2 Two Nvidia 3060 cards Cannot install the plugin and getting below error message --------------Can't download Nvidia Driver Package v590.48.01----------------- Removing package: nvidia-driver-2025.03.25 plugin: run failed: '/bin/bash' returned 1

I am able to access it over /mnt/cache/appdata, but not over /mnt/user/appdata. I am referring to "ls command" after SSH into the server.

Yes. Did disconnect and reconnect the usb cable before the diagnostic. Within 5min or so. If not software related, any idea how I can troubleshoot this issue?

Thanks. Changed it, but same issue.

I installed a second GPU and now I can no longer pass it through to the VM. I am getting the error message: "Start disabled due to PCI changes". Happens withever of the GPUs I select for passthrough. Please let me know if any thoughts or happy to provide logs. Thanks!

Thanks for your help and sorry for late reply. Unfortunately, I am still facing the same issue. I changed to a single-devices USB enclosure for easier problem-solving. Please see below diagnostic log. Thanks! tower-diagnostics-20260110-0749.zip

Below what I am getting when trying to see appdata as user share. Any idea what can be driving this. Config below. Can post diagnostic log if helpful. Please let me know. root@Tower:~# ls /mnt/user/appdata /mnt/user/appdata@ Exclusive access: Yes Primary storage: Cache Secondary storage: None Export: Yes Case-sensitive names: Auto Security: Public

Are ZFS pools yet supported?

Thanks for your help. Yes, route is approved. Amd don't think anything else is advertisig a similar route. In Unraid plugin settings, "accept routes = no", "accept dns = no", "allow tailscale subnets = no", "allow tailscale dns settings = no", and "tailscale funnel =no". I do believe the "accept routes" are turned on for the devices I am using remotely. But I don't think I am advertising any routes on them.

Got it. Here we go: ee94acfb6d2d0d07b1e11a1910a6e80f